habrproxy

proxy.py

import re
from http.server import HTTPServer, BaseHTTPRequestHandler
from urllib.parse import urljoin, urlparse

import requests
from lxml.html import tostring, document_fromstring


class MyProxy(BaseHTTPRequestHandler):
    to_host = 'http://habrahabr.ru'

    def text_replace(self, text):
        return re.sub(r'\b([^\W\d]{6})\b', r'\1™', text, flags=re.U)

    def do_GET(self):
        url = urljoin(self.to_host, self.path)
        response = requests.get(url, stream=True)
        output = response.content
        if 'text/html' in response.headers['content-type'].lower():
            try:
                page = document_fromstring(response.text)
            except Exception as e:
                self.send_error(500, e.message)
                return

            server_netloc = '{}:{}'.format(
                self.server.server_address[0],
                self.server.server_address[1]
            )
            links = '//*[@href]'
            for link in page.xpath(links):
                try:
                    ref = urlparse(link)._replace(netloc=server_netloc)
                    link.getparent().attrib['href'] = ref.geturl()
                except:
                    pass

            excluded_tags = ['script', 'style']
            for node in page.iter('*'):
                if node.tag in excluded_tags:
                    continue
                for text in node.xpath('./text()'):
                    clean_text = text.strip()
                    if clean_text and len(clean_text) < 6:
                        continue
                    parent = text.getparent()
                    if parent.tag in excluded_tags:
                        continue
                    if text.is_text:
                        parent.text = self.text_replace(text)
                    elif text.is_tail:
                        parent.tail = self.text_replace(text)
            output = tostring(page, encoding='utf-8')

        self.send_response(response.status_code)
        self.send_header("Content-Type", response.headers['content-type'])
        self.end_headers()
        self.wfile.write(output)


if __name__ == '__main__':
    httpd = HTTPServer(('127.0.0.1', 8000), MyProxy)
    try:
        httpd.serve_forever()
    except KeyboardInterrupt:
        pass
    httpd.server_close()